On the Interpretation of Lattice Diffusion in Magnesium Oxide

Viscosity-depth profile of the Earth's mantle: Effects of polymorphic phase transitions

“…However, the transition from intrinsic to extrinsic diffusion is not always obvious in oxides. For example, it has been argued by Wuensch [1975] that even at temperatures approaching T,,, pure intrinsic diffusion has never been observed in MgO. In view of such uncertainty it is also not obvious that diffusion-controlled creep in the mantle is ourely intrinsic; however, we will make this assumption.…”

“…and silicates in addition to some high-temperature creep data. The criterion Do >• I was used to select intrinsic data, although as is discussed above, this appears to be a necessary but not sufficient condition [Wuensch, 1975]. For example, even though Do = 2.04 ½m•/s for Fe•Os, we believe that the value of E0*, 79.9 kcal/mol, reported by Hagel [1966] does not represent intrinsic diffusion.…”

“…Nevertheless, we note that by choosing a smaller initial G* at a 320-km depth, say, 80 kcal/mol, activation volumes as low as 2 cma/mol could be obtained for the lower mantle, and our relatively high viscosity estimates could be brought into better agreement with the lower and more uniform viscosities inferred from glacial rebound studies. Such a low initial G* would imply that the diffusion which controls creep is not volume diffusion, or is not totally intrinsic, a real possibility in the mantle [Wuensch, 1975].…”

Diffusion in Oxides

“…Extrinsic anion vacancies are much less abundant in most samples than cation vacancies, because positively charged cation solutes (mainly trivalent cations such as Fe 3+ , Al 3+ and Cr 3+ ) tend to be present at higher concentrations than solutes with effective negative charges (monovalent cations such as Li + and Na + ). The commercial "nominally pure" MgO crystals used in most experimental diffusion studies are calculated to contain cation vacancies at the level of ~15-1000 ppm (e.g., Wuensch 1975;Oishi et al 1983), which would likely predominate over intrinsic cation vacancies at all temperatures up to the melting point.…”

“…On the other hand, when a small amount of Al was added to the tracer solution deposited on the crystals, the diffusion coefficients for Ni increased by a factor of ~4, indicating that diffusion was enhanced by cation vacancies introduced to charge-balance Al 3+ that diffused into the crystal during the experiment. Wuensch andVasilos (1962, 1971) also performed experiments using nominally pure MgO single crystals, with both vapor sources and NiO thin films used as the diffusant, and electron microprobe analysis of the diffusion profiles. Although these experiments were, strictly speaking, interdiffusion experiments, the NiO concentrations were small and the results similar to those from the Harding (1972) and Mimkes and Wuttig (1971) tracer studies.…”